#include <iostream>
#include <opencv2/core/core.hpp>
#include <opencv2/features2d/features2d.hpp>
#include <opencv2/highgui/highgui.hpp>
#include <opencv2/calib3d/calib3d.hpp>
#include <Eigen/Core>
#include <Eigen/Geometry>

#include <chrono>

#include <vector>

#include <ceres/ceres.h>

#include "PnPSolver.h"
#include "VerifyReprojectionJacobian.h"

using namespace std;
using namespace cv;
using namespace Eigen;

static double fx_, fy_, cx_, cy_;

template<typename T>
Eigen::Matrix<T, Eigen::Dynamic, Eigen::Dynamic> cvMatToEigen(const cv::Mat& cv_mat) {
    Eigen::Matrix<T, Eigen::Dynamic, Eigen::Dynamic> eigen_mat;
    eigen_mat.resize(cv_mat.rows, cv_mat.cols);
    
    for (int i = 0; i < cv_mat.rows; ++i) {
        for (int j = 0; j < cv_mat.cols; ++j) {
            eigen_mat(i, j) = cv_mat.at<T>(i, j);
        }
    }
    return eigen_mat;
}

void find_feature_matches (
    const Mat& img_1, const Mat& img_2,
    std::vector<KeyPoint>& keypoints_1,
    std::vector<KeyPoint>& keypoints_2,
    std::vector< DMatch >& matches );

// 像素坐标转相机归一化坐标
Point2d pixel2cam ( const Point2d& p, const Mat& K );


// 在优化后添加详细的结果分析
void AnalyzeResults(vector<Eigen::Vector3d> landmarks,
                    vector<Eigen::Vector2d> bearings,
                    const Eigen::Vector3d& estimated_translation,
                    const Eigen::Quaterniond& estimated_rotation) {
    
    std::cout << "\n=== Detailed Result Analysis ===" << std::endl;
    
    // 重投影误差分析
    double total_reprojection_error = 0.0;
    int point_count = landmarks.size();
    
    for (int i = 0; i < point_count; ++i) {
        Eigen::Vector3d p_c = estimated_rotation * landmarks[i] + estimated_translation;
        double u_pred = fx_ * (p_c[0] / p_c[2]) + cx_;
        double v_pred = fy_ * (p_c[1] / p_c[2]) + cy_;
        
        double error = std::sqrt(std::pow(u_pred - bearings[i][0], 2) + 
                                std::pow(v_pred - bearings[i][1], 2));
        total_reprojection_error += error;
    }
    
    double avg_reprojection_error = total_reprojection_error / point_count;
    std::cout << "Average reprojection error: " << avg_reprojection_error << " pixels" << std::endl;
}

int main ( int argc, char** argv )
{
    //-- 读取图像
    Mat img_1 = imread ( "../1.png", 1 );
    Mat img_2 = imread ( "../2.png", 1 );

    vector<KeyPoint> keypoints_1, keypoints_2;
    vector<DMatch> matches;
    find_feature_matches ( img_1, img_2, keypoints_1, keypoints_2, matches );
    cout<<"一共找到了"<<matches.size() <<"组匹配点"<<endl;

    // 建立3D点
    Mat d1 = imread ( "../1_depth.png", 1 );       // 深度图为16位无符号数，单通道图像
    Mat K = ( Mat_<double> ( 3,3 ) << 520.9, 0, 325.1, 0, 521.0, 249.7, 0, 0, 1 );

    fx_ = 520.9;
    fy_ = 521.0;
    cx_ = 325.1;
    cy_ = 249.7;

    vector<Point3f> pts_3d;
    vector<Point2f> pts_2d;
    for ( DMatch m:matches )
    {
        ushort d = d1.ptr<unsigned short> (int ( keypoints_1[m.queryIdx].pt.y )) [ int ( keypoints_1[m.queryIdx].pt.x ) ];
        if ( d == 0 )   // bad depth
            continue;
        float dd = d/5000.0;
        Point2d p1 = pixel2cam ( keypoints_1[m.queryIdx].pt, K );
        pts_3d.push_back ( Point3f ( p1.x*dd, p1.y*dd, dd ) );
        pts_2d.push_back ( keypoints_2[m.trainIdx].pt );
    }

    cout<<"3d-2d pairs: "<<pts_3d.size() <<endl;

    Mat r, t;
    solvePnP ( pts_3d, pts_2d, K, Mat(), r, t, false, SOLVEPNP_EPNP); // 调用OpenCV 的 PnP 求解，可选择EPNP，DLS等方法
    // solvePnP ( pts_3d, pts_2d, K, Mat(), r, t, false ); // 调用OpenCV 的 PnP 求解，可选择EPNP，DLS等方法

    Mat R;
    cv::Rodrigues ( r, R ); // r为旋转向量形式，用Rodrigues公式转换为矩阵

    cout<<"R="<<endl<<R<<endl;
    cout<<"t="<<endl<<t<<endl;
 
    vector<Eigen::Vector3d> landmarks;
    vector<Eigen::Vector2d> bearings;
    
    // 添加一些测试点
    for (int i = 0; i < pts_3d.size(); ++i) {
        landmarks.push_back(Eigen::Vector3d(pts_3d[i].x, pts_3d[i].y, pts_3d[i].z));
        Point2d points_2d = ( Point2d(pts_2d[i].x, pts_2d[i].y) );
        bearings.push_back(Eigen::Vector2d(points_2d.x, points_2d.y));
    }
    Vector3d eigen_translate = cvMatToEigen<double>(t);
    Matrix3d eigen_rotation_mat = cvMatToEigen<double>(R);

    // 初始位姿估计
    Eigen::Quaterniond initial_rotation(eigen_rotation_mat);
    Vector3d initial_translation(eigen_translate);
    
    if (!landmarks.empty()) {
        JacobianVerifier::VerifyReprojectionJacobian(
            landmarks[0], bearings[0], initial_translation, initial_rotation, fx_, fy_, cx_, cy_);
    }

    std::cout << " Initial translation ======== : " << initial_translation.transpose() << std::endl;
    std::cout << " Initial quaternion: " << initial_rotation.coeffs().transpose() << std::endl;



    // 创建求解器并求解
    PnPSolver solver(fx_, fy_, cx_, cy_);
    bool success = solver.Solve(landmarks, bearings, initial_translation, initial_rotation);

    if (success) {
        std::cout << "\n=== Optimization Result ===" << std::endl;
        std::cout << "Final translation: " << initial_translation.transpose() << std::endl;
        std::cout << "Final quaternion: " << initial_rotation.coeffs().transpose() << std::endl;

        AnalyzeResults(landmarks, bearings, initial_translation, initial_rotation);
    } else {
        std::cerr << "Optimization failed!" << std::endl;
        return -1;
    }
}

void find_feature_matches ( const Mat& img_1, const Mat& img_2,
                            std::vector<KeyPoint>& keypoints_1,
                            std::vector<KeyPoint>& keypoints_2,
                            std::vector< DMatch >& matches )
{
    //-- 初始化
    Mat descriptors_1, descriptors_2;
    // used in OpenCV3
    Ptr<FeatureDetector> detector = ORB::create();
    Ptr<DescriptorExtractor> descriptor = ORB::create();
    // use this if you are in OpenCV2
    // Ptr<FeatureDetector> detector = FeatureDetector::create ( "ORB" );
    // Ptr<DescriptorExtractor> descriptor = DescriptorExtractor::create ( "ORB" );
    Ptr<DescriptorMatcher> matcher  = DescriptorMatcher::create ( "BruteForce-Hamming" );
    //-- 第一步:检测 Oriented FAST 角点位置
    detector->detect ( img_1,keypoints_1 );
    detector->detect ( img_2,keypoints_2 );

    //-- 第二步:根据角点位置计算 BRIEF 描述子
    descriptor->compute ( img_1, keypoints_1, descriptors_1 );
    descriptor->compute ( img_2, keypoints_2, descriptors_2 );

    //-- 第三步:对两幅图像中的BRIEF描述子进行匹配，使用 Hamming 距离
    vector<DMatch> match;
    // BFMatcher matcher ( NORM_HAMMING );
    matcher->match ( descriptors_1, descriptors_2, match );

    //-- 第四步:匹配点对筛选
    double min_dist=10000, max_dist=0;

    //找出所有匹配之间的最小距离和最大距离, 即是最相似的和最不相似的两组点之间的距离
    for ( int i = 0; i < descriptors_1.rows; i++ )
    {
        double dist = match[i].distance;
        if ( dist < min_dist ) min_dist = dist;
        if ( dist > max_dist ) max_dist = dist;
    }

    printf ( "-- Max dist : %f \n", max_dist );
    printf ( "-- Min dist : %f \n", min_dist );

    //当描述子之间的距离大于两倍的最小距离时,即认为匹配有误.但有时候最小距离会非常小,设置一个经验值30作为下限.
    for ( int i = 0; i < descriptors_1.rows; i++ )
    {
        if ( match[i].distance <= max ( 2*min_dist, 30.0 ) )
        {
            matches.push_back ( match[i] );
        }
    }

    Mat img_matches;
    drawMatches( img_1, keypoints_1, img_2, keypoints_2, matches, img_matches );
    imshow("ORB Matches", img_matches);
    waitKey(0);
}

Point2d pixel2cam ( const Point2d& p, const Mat& K )
{
    return Point2d
           (
               ( p.x - K.at<double> ( 0,2 ) ) / K.at<double> ( 0,0 ),
               ( p.y - K.at<double> ( 1,2 ) ) / K.at<double> ( 1,1 )
           );
}